Continued from The Basics of (Mixing Part 1)

EQ
EQ is arguably the most important part of the mix. It is the practical and mathematical part of mixing. I helps me to try an visualize the frequency spectrum and come up with a general plan of where I want each track/instrument to sit. This is always a trade off with making things fit together and making each instrument sound good by itself. For example, acoustic guitars will often sound unnaturally boomy without cutting some frequencies out around 200 Hz, this also makes some room for the bass guitar and sometimes the toms of the drum kit. This is a balancing act that is a whole new monster with each mix so all of the above are very loose guidelines. The purpose and goals of the EQ stage are three fold: to establish a good balance, to enhance the frequencies that you like in each track, and take away the ones that detract from the sound, and to eliminate any masking of one instrument over another. Here is a post dedicated to EQ.

Compression (dynamics)
Compression also has multiple uses in the mixing environment, it has uses as a tool, and also as an effect. Compression is often used as a tool to control the dynamic content of a track. In this scenario I am usually trying to control stray loud or soft content and regulate them so that they are approximately the same volume as the other surrounding content. Note that not every track will require compression, if the performance is consistent (which is more often than not the case with well practiced musicians) then compression may not be needed in this capacity. As a general rule if it doesn’t need compression then don’t apply any, at least not locally. On a side note I often will mix with subtle compression on the mix bus at all times, I find that I require less compression on each individual track and gain some consistency between all the tracks. (Maybe there is some truth to the statement that the SSL quad compressor is like mix glue)

Compression can also be used as an effect, but I find when used this way it is best used in a send/return format. (This is often referred to as parallel compression) I will often send the close mics of my drum kit to a bus where I will compress the heck out of them, with not too fast of an attack, because the purpose of this compression is to make the drums punchier I want to keep the transients. This sound on it’s own is not that great to listen to on it’s own, but bring it up slowly in the mix and blend it with the original close mics for a punchier and edgier sound. I also have a post that explores compression in more detail.

Panning
Now that I have most of the frequency and dynamic content spaced out and balanced I find I need to add some width to the recording. I will apply panning to the tracks at this point usually starting with the drum kit. Panning the toms can make the kit sound bigger/wider, but I usually don’t pan them hard left and right as that can be a little distracting on tom rolls. If you used spaced pair overheads or as room mics on the kit I would pan them according to the same principals. As a general rule I will pan instruments with similar frequency content to either side such as an electric guitar slightly left, and acoustic guitar slightly right. If you’ve seen the band play live visualize where each player stood as this can be a natural panning image to try and recreate. The human ear loses the ability to place sounds below a certain frequency and above another so panning will become ineffective. The effective frequency range for panning is between (150Hz and 1.5kHz). For content above 2.5kHz, if you have a stereo signal to begin with, differences in EQ between each channel that will create phase effects that will place sounds off center in the stereo image. For content below 150Hz, if you have a stereo signal to begin with, you can use the Haas panner (precedence) effect by delaying one of the signal by a very small amount (0.2-0.44 msec). Once your stereo image is in place, the practical side of mixing is essentially complete and the more artistic and creative part is applied. These are the techniques that will make your mixes unique and stand out.

Send Effects (Reverbs and Delays)
The use of effects is very much up to personal taste or the taste of the band that you are mixing for. This part of the mix should be very much inspired, one of the signature sounds of an amateur mix is the misuse or overuse of effects. Remember your job as the mix engineer is not to define the sound of the band, but to support and enhance the music that the band has produced. It is OK to take some ownership of the music, but ultimately it is the bands/clients music. Even if you are mixing your own bands music it helps to treat the mix process this way. This stage of the mix is where you will define yourself as an engineer, so I cannot tell you exactly what to do, just offer some tips and guidelines. Reverb and delay can be used effectively to ad depth to a mix. Listen to the drum kit with your eyes closed as you add a big long reverb as you increase the reverb the kit should start to sound like it is farther away. You can use this principal on all of the tracks to give the mix depth. You probably also noticed with the drum kit that as you increased the reverb the cymbals became really splashy and wishy-washy. Often it helps to EQ and filter the signal that you send to the reverb. I typically will band pass filter my reverb and delay sends because to much top gets splashy and too much bottom gets muddy. Many of the new convolution reverb plug-ins give you tremendous control over the reverb tail and decay times this is where you should reach if you want to make your reverb sound more natural or more out of this world. I tend to like plate reverbs and short reverbs for vocals as you the can add a reverb effect to the voice with out moving it from front and center. Slapback reverbs can be nice to really get a crackin’ snare. Really though all of these things depend on the style of music, and the sound that the band is going for. For more on finishing up a mix please see The Basics of Mixing (Part 3).

There are no hard and fast rules when it comes to mixing or any part of the music production chain for that matter. Everything mentioned in this series is meant to be a guideline or a suggestion, and absolutely not the only way to do things. This is will series will follow the basic flow that I go through when a dig in to mix a track. It is also worth pointing out that what I consider mixing is the process of blending the sounds that are to be used in the song together, therefore I typically will have all tracks recorded, all overdubs and BV’s tracked, and all editing done. It doesn’t always work out like this, but I find that if I can take care of as many of those things as possible before I sit down to mix the whole process of mixing is more streamlined and inspired. I will also assume that this mixing session is to be conducted in the box on a computer as this will apply to most people reading this post. (There are some significant differences when mixing on an analog console from a media such as tape.)

Gain structure
The first thing that I do when mixing is set up the gain structure. In the digital world ideally this would have been taken into consideration when tracking. When tracking digitally and mixing in the box there is a significant chunk of the typical signal flow missing from the signal chain. It is the preamp from the tape machine back into the mixing desk. This is not available when mixing in the box. You are very much at the mercy of you record levels. Getting good signal to noise ratios is important (both in the digital and analog realms), but when mixing digitally you don’t want to record too hot either. Firstly, because digital clipping sounds awful, and secondly, you don’t have that extra preamp when setting the gain structure. When tracking digitally I try to get the sustained levels of each track to be around -3dB. This will give me (and eventually the mastering engineer) a little more head-room to work with. Notice I said each track it is always tempting to mix a little bit with the record levels when tracking, but would suggest trying to level the playing (volume) field for all tracks when tracking.

I know I said you don’t start mixing until everything else is done, but good and consistent record levels really are the first step of mixing, you need a foundation on which to build. Setting a gain structure was typically done with the missing preamp in the digital audio chain, which is why I mentioned getting good record levels. Try and get all the tracks to equal volume (this most likely will not sound good). The reason I stressed record levels is because normalize and gain plug-ins are the bullies of digital audio (especially the ones that re-render wave files) an have a very high potential of mangling your pristinely recorded tracks. If you didn’t record at all equal volume I would suggest using the next step to make up any gain on any track that you are missing.

Filtering
The next step when setting up a mix is filtering. During this process I toggle solo on the tracks often so becoming familiar with this function would be a good idea. I work my way through the tracks filtering out the top and bottom of each track so that it contains only the sounds that I want from the track. For instance an electric guitar an amp rarely produce any musical content below 80 Hz so I will set a high pass filter around 80 Hz on all the guitar tracks. People often ask “how come I can’t get my track as loud as all the others out there, even when I limit and compress the hell out of it.” Often times they have not taken the time to take out the noise that resides at the top end and bottom end of every track. We can’t even hear this noise sometimes, but electronically or digitally it is eating up our head-room, so take the time to knock it out, you’ll thank yourself when you enter the “loudness wars.” Another thing to think about when doing this is exactly what sound am I trying to get out of this track. For example often time the drum kit comes in on several tracks and suppose there is a close mic on the hi-hat as well as a mic on the snare and some overhead and or room mics. That hi-hat track can often be pared way down in the frequency spectrum as you will be able to collect the missing parts of the signal from the rest of the mics. Another rule of thumb, if you can’t figure out what sound you are trying to get out of a track, don’t use it, it is just eating up space in your mix.

Signal Routing
Once I have the gain structure and filtering under control I move on the signal routing. In Pro Tools I will set up all the sends for my send effects, I will set up my virtual stereo master bus, and any other sends or auxiliaries that I plan to use. I will set up a big or long reverb send, a medium one, and a short one as well as a plate (four different sends). I will also setup at least a couple of delays: one short and one long, and any other time based effects like chorus or flanging effects with their own send. This may look like you are adding a lot of tracks, but by setting them up as sends you can share the effect across many of your recorded tracks, which will actually save you computer power in the long run. I also often like to parallel compress the close mics of the drum kit so I will setup a send for them as well. Remember that sends can be taken pre fader or post fader. A pre-fader will send the signal straight from the recorded signal with no effects applied, a post fader send will send the signal after it has been processed by any effects in the channel.

At this point I am ready to set up my basic mix, I will pass through the faders and get a good balance between the tracks making sure to leave about 5 dB or head-room for when I return my send signals back to the master stereo bus and to leave some room for the mastering engineer to work with. (Continued in The Basic Mixing (Part 2))

Sequencing is the process of programming a performance into a computer program to allow the computer to play the performance back to you. In this way a single individual can make music by programming the computer to play multiple performances simultaneously. Today this is typically done through a protocol called MIDI (Musical Instrument Digital Interface). MIDI is most commonly displayed in a Piano Roll format that shows when notes (Events) occur vertically, and time horizontally. Sometimes this information is also displayed in a list, but most musicians find the Piano Roll more comfortable, there are even some people that the “Piano Roll” display may replace traditional music notation. There are three basic ways to accomplish sequencing.

The first is to record a real-time performance from a MIDI controller. MIDI controllers are most commonly keyboards, but there are many new interfaces such as guitars and flutes. These controllers send performance data (not actual sound) to the computer or sound card, which the computer then records and stores against time. Essentially the computer is able to replicate the performance, giving the composer the ability to tweak and adjust the sound(s) associated with the performance. This is often the most efficient way to input a sequence.

The second way to accomplish sequencing is by manually placing notes on the timeline. Most software packages offer the ability to manually enter notes (and other performance data) directly on the time-line, this is usually using a “Pencil” tool. In this way a composer can write music that he/she may not have been able to perform otherwise. This procedure does not require a MIDI interface and is available to most composers out of the box with their recording application.

The final way of accomplishing sequencing is through a process often referred to “Step Input.” Step Input involves entering the performance using a MIDI controller, but not in real-time. The composer is able to enter notes one at a time and often use auxiliary buttons or keys on the controller to advance through the time line. As an example if the composer wanted to enter a C major triad, he/she could press the C key, then the E key then the G key, and then press the button to advance to the next time slot. Often other auxiliary buttons are also available to change note durations. This method takes some time to get used to, but down the road can be an efficient way to input a sequence.

MIDI offers much more data than just note placement and duration. MIDI data can be assigned to control virtually any parameter of a sound, but some of the most common are the Velocity, Sustain, and Pan. Velocity is a measurement of how hard a note is struck and can have a large effect on the timbre of the sound. Think about how different a guitar sting sounds when picked softly or very hard, there are differences in the sound other than just the volume, velocity tries to emulate these performance differences. All of this data is then fed to a sound generator.

The sound generator (typically a virtual instrument, sampler, or synthesizer) interprets this data and creates an audio signal. One of the advantages of capturing a performance via midi rather than acoustically is that you can apply that performance to any sound generator you would like to. In this way a keyboard performance can sound like drums, or guitar, or french horn.

The major drawback of MIDI is that it rarely can sound as authentic as a real acoustic performance. This is due to the fact that rich sample libraries cost a lot to produce, and can take an infinite amount of space depending on how accurate you would like to make them. MIDI also suffers from some latency and a limit to how much data can be transferred in “Real time.” MIDI data is transmitted serially (one 0 or 1 at a time), the transmit speed is only, 31.25 Kilobits per second (over a traditional MIDI 5 pin connection). This is fast enough for us not to hear the small amount of time that resides between the sounding of each note in a chord, but if several performances are being transmitted to several sound generators via the same MIDI connection some of these timing issues become recognizable to human perception.

That’s the basics, and the major advantage and disadvantage of MIDI and sequencing. For detail on how to implement MIDI in your home studio refer to your owners manuals for your gear.

I was recently glancing over some of the articles popping up on home recording blogs and came across this so called “Recorderman” overhead drum mic placement technique. A good technique built on sound principles. This brings to light some interesting thoughts. This technique can be expanded and applied to any multiple microphone situations. The process of keeping both mics equal distances from the main sound source (in the given example the kick and snare) eliminates what is commonly referred to as phasing.

Phasing occurs when the same sound arrives at two different mics at different times. Sound travels at a particular speed, but always at the same speed (in the same medium, in most cases in the studio: Air). Therefore if the microphones are the same distance from the source and point at generally the same part of the source, they should yield a similar sound (especially if they are a matched pair).

When used in this way multiple microphone techniques can yield good “images” of a sound source with minimal phase issues. The “image”, referring to the stereo placement of sounds, is the sound coming from the left or right side. This also is the logic behind coincident mic techniques i.e. M+S and XY where the mic capsules are placed as close together as possible (which in turn makes them equal distances from the source).

The downside of phasing and the use of multiple microphones is a phenomenon known as comb filtering. When out of phase signals are summed (mixed together) parts of each wave form will cancel, resulting in silence. This is not a general silence but the elimination of specific frequency depending on how out of phase the signals are. Frequencies very near the canceled frequency will also be affected to less and less of a degree as you move away from the canceled frequency. This is very much like applying a notch filter to a sound to eliminate a particular frequency. To make matters worse this notch effect will also occur at each harmonic of the canceled frequency. For example, if 800 Hz is canceled, then 1600, 3200, 6400, and 12800 Hz will also be affected. If you can imagine an EQ with a notch filter on each of the above frequencies you should get an idea of why they call it “comb” filtering. This effect can be simulated and sometimes overlooked when using a delay plug-in. Sweeping the delay time of a delay plug-in will sound similar to sweeping a notch filter up and down the frequency spectrum. Try it!

In conclusion, may I offer two additonal pearls wisdom? 1. Don’t use more mics than necessary. 2. If you want to use more than one mic to create a stereo image be very intentional about how you do it. (A tailors cloth tape measure can be very good investment in a studio.)

Continued from: Mixing Digitally - Adding Warmth to Clarity (Part 1 - Background)

The first and most logical step to adding warmth to Digital recordings is to use nice analog gear before the A/D converter in the recording chain. Using mic-pres and tube gear can be great ways to introduce some warmth and color to the signal when recording. Always keep in mind though that any color you add at this point is printed and cannot be removed (at least not easily). Recording through a SSL channel strip (or the whole console if you’ve got one) can produce some very nice results.

You might be saying “Well come on Ian, not everyone has a SSL to record through.” You’re right, so here’s the plan. Most of us only have whatever pre-amps our sound card came with and typically the plug-ins that came with our application. We need to create a system to put the same distortion or color into our mixes as the analog gear of yesterday.

Step 1: Get a sense of how the old gear and your favorite recordings sound. Are the mixes bright, or punchy, how do the drums sound, what kinds of effects were used? It’s always easier to get the sound you want when you know what the sound you want sounds like :).

Step 2: Setup a virtual mix bus. As you do more and more mixing you should get a feel for how you like your master bus to sound. As this becomes clear try to establish some consistency. Just like using an analog mix bus that has the same circuitry for every mix so should your digital mix bus. Do this first before you set mix levels or anything. Come up with a set of plug-ins that will be your custom mix bus and and mix the entire time through them. McDSP makes some excellent plug-ins for this but look for plug-ins that have channel characteristics (i.e. EQ and Dynamics). Avoid multi-band compressors at this point though; leave that for the mastering stages. For example, my mix bus on my Pro Tools LE system at home always has a McDSP “Analog Channel” and a P4 equalizer.

*Note: If possible, don’t stereo link the plug-ins on the mix bus and tweak the settings on each just by the littlest bit so they don’t match exactly. This can create a more natural sounding stereo image. (Remember on an analog console no two channel paths are exactly the same due to the manufacturing tolerances of electrical components.)

Step 3: If you have the DSP power, extend this principal to each individual track applying EQ and Dynamics to each channel as if you were using the inline modules of a real mixing console. This is also where I would encourage you to apply the “Tape Saturation” or “Tube warming” plug-ins if you’ve got them. Remember to use these effects for subtle coloration of the sound. If you’ve got the power, I like to EQ before and after the Saturation and/or warming plug-ins. Remember, we are emulating the analog signal flow back from a real tape machine.

Step 4: This may or may not be considered a warmth issue but it does kind of apply as far as emulating outboard signal flow with plug-ins. When using sends for effects like reverb (you are using sends or aux’s right :)), if possible, EQ the send for certain signals. Don’t send too much top end (cymbals) to the reverb as this can produce a very splashy sound in high frequencies. Don’t send too much of the bottom end either as this can very quickly muddy up your mix. It’s typically best to band pass your reverb sends and if get really creative you can EQ all your reverb signals so that they sit well together. Often Outboard gear, because of cable runs, inherently had some bandpass filtering and many times the circuitry would limit the frequency spectrum, as well.

Using this setup will also force you to use the same plug-ins consistently thus give you a firm understanding of their use. This will make you much more efficient as an engineer and give you the confidence that you know how to achieve the sound you are looking (listening) for.